is 40mp too good for lenses?

With a 3.1 um sensor, that would be f/9.

Every Fuji GF lens gets over that bar.

There are three considerations that are relevant here.

The upper spatial frequencies in the subject. If the subject has no content above Nyquist, you won’t see aliasing.

Technique. Anything that causes sufficient blur will suppress aliasing.

The viewer. Lab targets to detect aliasing are chosen to make artifacts obvious. The viewer of an image of a natural scene may not know what the real scene looked like.

In addition, some people seem to like aliasing. Examples are some pixel peepers and those who like the look of MF fat pixel sensors. I remember people talking about how much they liked the way that fat pixel sensors made skin sparkle.

There is aliasing, I'm sure of it. But my example was not so good. It is a cropped screenshot of a picture shown at 200% in Gimp. I think here it is magnified more, maybe to 250 - 300%. So, it is pretty soft, when at the same time the lens is too sharp. 😅

I post here a crop from the opened picture, a little sharpened.

Jim,

I think we are totally at cross purposes here, probably due to lack of clarity on my part. My comment was not about aliasing, but about the three conditions that you laid down in your statement:

I replied:

You have responded to the "decent" criterion by apparently limiting it to Fuji lenses. What I meant by the other two are that "on axis" is just the centre of the photo, and "optimal aperture", is also specific for a given lens.

In practice, the user wants to be able to use the lens at other than the optimal aperture, and hopes for good (i.e. reasonably uniform) performance over the whole image consistent with the manufacturer's claims. This was the basis of Alan's complaint (the image is softer on one side than the other).He was not referring to the optical axis, or restricting his comments to f/5.6.

I hope this resolves the apparent disagreement! :)

Best,

David

I don’t know why you are thinking that I am defining decent as GF.

I used that example since I have tested all the GF lenses. I’ve tested a lot of Zeiss, Sony, SK, CV, Rodenstock, Nikon Z, and other lenses. I’ve hardly ever seen one that couldn’t meet that test.

True enough. The user also wants the lens and camera to deliver good performance under optimal conditions. Over sampling some portions of the image doesn’t hurt any thing.

This is on the issue of 40mp being too good, not the issue of a sample defect.

Hard to tell with processing and downsizing. Is the raw available?

Back in my Kodak 14n days, over on the DPR Kodak forum we used to call those lovely colour sprinkles on the hairs on that leaf "Christmas tree lights". They were every where on that camera unless you blurred the a & b channels in LAB mode.

I once took a picture of a wet asphalt driveway with the 14n and it looked like an oil slick.